The paper introduces Multiscale Combinatorial Grouping (MCG), a unified approach for bottom-up hierarchical image segmentation and object candidate generation. The authors develop a fast normalized cuts algorithm to efficiently compute eigenvectors, a high-performance hierarchical segmenter that leverages multiscale information, and a grouping strategy that combines multiscale regions into accurate object candidates by exploring their combinatorial space. Extensive experiments on the BSDS500 and PASCAL 2012 datasets demonstrate that MCG produces state-of-the-art contours, hierarchical regions, and object candidates. The key contributions include an efficient normalized cuts algorithm, a state-of-the-art hierarchical segmenter, and a grouping algorithm that explores the combinatorial space of multiscale regions. The approach is flexible and can adapt to specific applications and object types, producing candidates at any trade-off between number and accuracy.The paper introduces Multiscale Combinatorial Grouping (MCG), a unified approach for bottom-up hierarchical image segmentation and object candidate generation. The authors develop a fast normalized cuts algorithm to efficiently compute eigenvectors, a high-performance hierarchical segmenter that leverages multiscale information, and a grouping strategy that combines multiscale regions into accurate object candidates by exploring their combinatorial space. Extensive experiments on the BSDS500 and PASCAL 2012 datasets demonstrate that MCG produces state-of-the-art contours, hierarchical regions, and object candidates. The key contributions include an efficient normalized cuts algorithm, a state-of-the-art hierarchical segmenter, and a grouping algorithm that explores the combinatorial space of multiscale regions. The approach is flexible and can adapt to specific applications and object types, producing candidates at any trade-off between number and accuracy.